Multi-Agent Systems

As agentic coding systems decompose work across multiple model instances, a critical safety question is whether those instances can coordinate to achieve a hidden malicious objective while remaining aligned with user intent. We introduce SCHEME, a benchmark of 17 task instances across 7 settings and 8 real open-source libraries, each pairing a legitimate software-engineering task with a covert side task. Every setting is designed so that no proper subset of agents can succeed alone: agents must decompose a shared sabotage plan, relay partial requirements under different communication topologies, and execute mutually consistent edits, testing genuine multi-agent coordination rather than individual capability. Evaluating with GPT 5.1 Codex and Gemini 3.1 Pro, we find coordinated sabotage is already practical, with Gemini completing the covert objective while succeeding on the legitimate task in 84% of samples and Codex in 46%. A failure analysis shows that recovery, not failure incidence, drives the model gap: both models suffer similar numbers of communication failures per trajectory, but Gemini recovers from $\sim$73% of occurrences against Codex's $\sim$56%. However, the agents' code edits remain detectable: a trusted monitor flags coordinated sabotage at FPR$\le$2% in 99%/68% (Gemini/Codex) of attempts from code edits alone, rising to 100%/81% with access to agent communication. Despite stealth-aware prompting, undetected sabotage stays at $<$3% under the full-visibility monitor; telling agents a monitor is present halves Gemini's sabotage success without significantly breaking detection, even when agents are explicitly prompted to fabricate cover stories.

Medication recommendation predicts medications for patient visits, but existing methods still face two key challenges. At the model level, traditional drug recommendation methods only predict structured drug codes with limited evidence grounding, while LLM agents can use richer clinical context but may lack safety verification and traceability. At the task level, existing benchmarks often use broad medication categories, which ignore subgroup-level safety differences and can lead to risk overestimation. We introduce the first fine-grained medication recommendation setting based on fourth-level ATC code generation. We propose Safe Prescription Agent (SafeRx-Agent), a knowledge-grounded multi-agent framework that uses patient context, external clinical knowledge, and safety verification to recommend traceable medication sets. Experimental results on MIMIC-III and MIMIC-IV datasets show that SafeRx-Agent improves fine-grained medication prediction accuracy while controlling drug interactions, contraindications, and medication set size.

We introduce the incremental voter model (IVM), a discrete-opinion multi-agent system where agents undergo step-wise transitions biased by the opinion of a randomly selected persuader. Our incremental voter model comprises a large population of interacting agents, each holding an opinion represented by an element of the discrete set $\{-k,\ldots,0,\ldots,k\}, k \in \mathbb{N}_{+}$. At each update step as time progresses, a pair of distinct agents are selected independently and uniformly at random from the population, and the first agent (viewed as the listener'') updates its opinion based on that of the second (viewed as the persuader''), adopting a new opinion that differs from its current one by at most one unit. By deriving the mean-field system of nonlinear ordinary differential equations (ODEs) that governs the large-population limit of the agent-based model, we develop a rigorous mathematical framework to study the asymptotic behavior of the opinion distribution in the mean-field limit. These results contribute to a deeper understanding of social influence processes in complex systems, particularly in modeling opinion polarization, and may guide the formulation of more advanced models in future research.

Role-based LLM multi-agent systems need adaptive role pools, yet adapting such systems is not merely a matter of prompt optimization: roles often carry structural obligations, including capability coverage, message compatibility, validation, final-answer aggregation, and parser-compatible output protocols. Existing systems either fix the role inventory and lose adaptivity, or allow unconstrained generation to induce role drift, removing structurally necessary roles and breaking answer contracts. We formulate this as contract-preserving role evolution, requiring every committed edit to preserve five structural contracts (capability, communication, validation, aggregation, output protocol). We instantiate this formulation in SERO, a Self-Evolving Role Orchestration framework that evolves a typed role-card pool through credit-guided retrieval, a credit-ranked communication DAG with a protected terminal aggregator and conditional validator repair, and a contextual-bandit controller whose LLM-proposed edits are committed only when they preserve the contracts and improve task score. Experiments on real-world reasoning benchmarks across three LLM backbones confirm the value of contract-preserving role evolution.

Latent-based multi-agent systems replace parts of explicit inter-agent communication with hidden representations, offering a new direction for efficient and flexible agent collaboration. However, moving coordination into latent space may also move attacks beyond the reach of visible-text inspection. In this paper, we study whether latent states can carry attack-associated information that remains effective during clean executions. To examine this question, we introduce a latent attack framework that reactivates attack-induced effects through latent interventions without reusing adversarial text. Extensive experiments show that the resulting latent-only attacks can substantially degrade task performance in clean executions, especially when applied to inter-agent KV-cache handoffs rather than local hidden states. Further control analyses indicate that this degradation cannot be reduced to arbitrary perturbations or invalid generation. Overall, our findings suggest that latent-based collaboration does not remove attack risk. It shifts part of the risk into less observable execution states, calling for safeguards beyond visible-text inspection.

Recent years have witnessed the rapid development of Large Language Model-based Multi-Agent Systems (MAS), which excel at collaborative decision-making and complex problem-solving. However, malicious agents in MAS may inject misinformation to mislead other agents and disrupt system performance, giving rise to a new research direction that focuses on attack mechanisms and defense strategies in MAS. Prior studies largely assume malicious agents act independently and investigate the corresponding defense strategies. However, we argue that malicious agents may exhibit collaborative behaviors, enabling more effective attacks through internal information exchange. In this paper, we propose an adaptive cooperative attack framework, where malicious agents autonomously coordinate and dynamically adjust their attack strategies through multi-round interactions. Furthermore, we introduce Sentence-Level Trustworthiness Analysis and Rectification (STAR), a defense framework that identifies and rectifies misleading information at the sentence level within agent communications. Our experiments show that cooperative attacks lead to a significantly larger degradation in task success rate than independent attacks, resulting in a relative drop of 5.34%. Meanwhile, STAR effectively mitigates both cooperative and independent threats and improves task success rate by an average of 36.76%. The code is available at https://github.com/smoooom/STAR.

Multi-agent systems are increasingly deployed to support various tasks where agents interact to achieve individual and collective objectives. Although these systems can enhance task performance and decision-making, fairness preservation through bias reduction remains challenging. This study examines how agent-level biases shift and impact system-wide fairness. We use prompts to expose individual agents to group-favoring bias, then assess downstream impacts at the system level. To quantify the impact, we propose Favor Bias Strength (FBS), a zero-centered metric that decomposes bias alteration between favored-group uplift and disfavored-group suppression. Using multiple agent designs, benchmarks, and up-to-date large language models, we show that agents endowed with bias can substantially affect system-wide fairness. Interestingly, when agents are exposed to bias uniformly, the system-wide bias elevates, even exceeding the additive sum of the individual agents' biases. The empirical evidence underscores the criticality of fairness in multi-agent systems, which warrants further analyses and empirical tests.

Effective multi-agent systems cannot be designed by selecting prompts or communication graphs in isolation. Agent behavior depends on the information an agent receives, while the usefulness of a communication edge depends on how the receiving agent interprets and uses that information. We propose \textbf{TCP-MCP} (Topology-Coupled Prompting for Multi-Agent Collaborative Problem-Solving), a co-evolution framework that searches agent prompts and communication topologies as a unified genome. TCP-MCP uses an initialization-time landscape probe to calibrate early search behavior, and then relies on Pareto-front diagnostics to adapt exploration under three objectives: task performance, token cost, and structural complexity. Using the same DeepSeek-V3.2 backbone across all methods, TCP-MCP achieves 82.66%, 89.96%, and 96.61% accuracy on MMLU-Pro, MMLU, and GSM8K, respectively. Across the three benchmarks, it consistently outperforms automated graph-generation baselines and achieves competitive accuracy relative to debate-style systems, while using up to 5.69$\times$ fewer tokens than those systems at the reported operating points. These results show that jointly evolving prompts and communication structure provides a practical route to cost-aware and task-adaptive multi-agent system design in controlled evaluations.

Multi-agent systems (MAS) have substantially advanced autonomous software engineering (SWE), but their growing inference energy demands raise sustainability concerns. In this paper, we demonstrate that this cost is concentrated in an overlooked source: redundant output tokens generated across agents. Two empirical findings ground this claim. First, our per-token energy attribution for MAS reveals a sharp asymmetry: an output token consumes 30 to 1,000 times more energy than an input or cached token. Second, MAS inflate per-episode output because agents repeatedly re-explore overlapping repository regions. To address this inefficiency, we propose Librarian, a persistent search sub-agent that tracks repository-search history and suppresses redundant exploration actions across agents. By returning short references to file regions instead of full file excerpts, Librarian further reduces output-token volume. On SWE-Bench Verified and SWE-Bench-Live, Librarian reduces per-episode GPU energy consumption of existing multi-agent SWE systems by 11 to 30% while preserving task performance. Our code is available at https://github.com/ml-postech/Librarian.

As multi-agent systems (MAS) become increasingly complex, identifying the contributions of individual agents is critical for system optimization. However, existing approaches lack a rigorous, unified framework for credit assignment. In this work, we formalize agent attribution as a cooperative game, parameterized by the coalition distribution, removal protocol, and target metric. Using this framework, we show that Leave-One-Out (LOO) identifies bottleneck agents as effectively as combinatorial methods, but at a fraction of the computational cost. We also demonstrate that removal protocols induce distinct games: Agent ablation isolates structural bottlenecks, whereas introspective LLM judges fail to faithfully approximate this behavior. Furthermore, to evaluate the utility of specific agent backbones, we introduce attribution via model replacement. By substituting underlying models of low-contribution agents, we improve task performance by up to 17% while reducing cost by up to 35% across three benchmarks. Finally, we apply our framework to audit a medical MAS, revealing that agent contributions to diagnostic accuracy and ethical behavior are often decoupled. By intervening on counterproductive roles, we observe an increase in ethics alignment while maintaining diagnostic accuracy. Overall, this work provides a principled approach for cost-effective MAS attribution and intervention.

Ensuring agent behaviors in distributed open multi-agent systems remains challenging, especially as populations grow and unaligned agents may exist. We show that a single aligned agent can propagate cooperative behaviors to untrained agents purely through natural language interaction, a phenomenon we term Alignment Propagation. We study this in the Red-Black Game, a team-based iterated Prisoner's Dilemma in which teammates deliberate and vote to determine their team's collective action. By distilling the cooperative reasoning and persuasive dialogues of a teacher model into a Qwen-3-14B, we obtain a seed agent that, when placed among four untrained teammates, doubles the cooperation rate from 24.8% to 62.2%, outperforming the teacher model and a vanilla Gemini-3.1-Pro. Remarkably, a seed trained exclusively on the RedBlack Game transfers zero-shot to Sugarscape, a spatially grounded survival simulation with pairwise trading, achieving a 91.5% trade success rate versus a 21.6% baseline. Our results reframe multi-agent alignment from an exhaustive per-agent training problem to a scalable social capability that can be engineered through strategic seed placement.

Modern analytics systems are fundamentally reactive, requiring users to define queries over increasingly complex and continuously evolving data. In real-time streaming environments, this paradigm breaks down, as the space of potential insights becomes too large to enumerate manually. We present a multi-agent architecture for autonomous insight discovery over real-time data streams. The system implements a continuous discovery loop in which agents generate hypotheses, compile them into executable analytics, validate generated artifacts, and produce visualizations and deployable applications. The architecture leverages Apache Kafka for event-driven coordination, Apache Flink for stream processing, and large language models to implement specialized agents. A key contribution is a contract-driven design based on typed intermediate artifacts, enabling modularity, observability, lineage, and safer execution of dynamically generated analytics. Through use cases in retail, finance, and public data, we show how this architecture supports a shift from query-driven analytics to proactive, discovery-driven systems.

The Brazilian Equatorial Margin (BEM) is Brazil's next offshore oil frontier, with operations expected to begin in 2026 in the Foz do Amazonas basin. Its assets are fiscally and territorially linked primarily to Maranhao -- the state with the lowest HDI in the Federation (0.676, IBGE 2022). This raises the central policy question: under what conditions does BEM exploration generate net positive externalities for Maranhao? The problem is intrinsically multi-agent: the Federal Government seeks revenue and energy security; the state seeks regional welfare under constitutional royalty earmarking; the operator maximizes profit under risk; ANP and IBAMA hold conflicting mandates; and Amazonian communities prioritize territorial and environmental vectors over monetary income. We present Margin Play, a Multi-Agent Reinforcement Learning (MARL) system simulating these tensions under Brazilian empirical calibration and classical economic literature. It implements six agents under the CTDE paradigm, trained with BRO-MARL. Results from 60,000 episodes across six scenarios indicate the answer is conditional on the institutional regime: under the reference baseline, the welfare gain is marginal (Waval approx. 1.68), whereas the MA-Prospero configuration yields Delta W = +17.5% and Delta Rcom = +21.3%, with a lower environmental liability (Eamb = 0.048 vs. 0.076). The fundamental problem is not a trade-off between production and welfare, but the choice of public policy regime linked to exploration.

Multi-agent systems built on large language models (LLMs) require many coordination choices that are difficult to fix a priori: which skill protocol to invoke, which agent role should perform a subtask, which model to bind to each role, how roles should interact, when to use retrieval or verification, and when to omit a step entirely. These choices interact with task regime and operational constraints, so static pipelines and one-off model comparisons provide only a limited view of the design space. This paper introduces AgensFlow, an open-source framework that treats multi-agent coordination as an online policy-learning problem under partial observability. The framework makes coordination decisions observable and learnable from repeated trajectories, rather than treating skill, role, model, topology, and evaluation choices as fixed pipeline design. AgensFlow is evaluated on two corpora: distributed-systems incident tasks and security-advisory tasks. The evaluation shows three main results: learned routing reaches a higher-quality operating point than a fixed pipeline baseline on coordination-heavy classes; skip:X isolates topology compression as a meaningful part of the substrate; and warm-started policy graphs can reduce exploration cost while preserving plateau quality. Overall, the results support that learned, auditable routing can improve coordination-heavy multi-agent workflows over static wiring.

LLM-based multi-agent systems decompose complex tasks into interacting roles, but most remain manually orchestrated by prompts, tools, and control rules, while agents are rarely optimized through a unified reinforcement learning interface. Existing RL post-training frameworks mainly target single-policy optimization and lack abstractions for user-defined multi-agent workflows, structured interaction, role-specific credit assignment, and configurable parameter sharing. We present UnityMAS-O, a general RL optimization framework for LLM-based multi-agent systems. UnityMAS-O treats the complete workflow as the optimization unit, rather than a single response or policy trajectory. It represents workflows through four first-class objects: logical agent roles, graph trajectories, user-defined rewards, and agent--model mappings. This decouples logical agents from physical model parameters, supporting full sharing, full separation, and partial sharing, with rewards assigned at role, turn, and trajectory levels. UnityMAS-O extends verl with a Ray-based star-topology runtime. A central controller executes workflows, invokes tools, records structured trajectories, and assembles rewards; model-local worker groups handle rollout, buffering, advantage computation, and distributed PPO-style updates. Users can define agents, workflows, model mappings, and rewards without rewriting the optimization infrastructure. We instantiate UnityMAS-O on retrieval-augmented QA, iterative agentic search, and reflective code generation. Across Natural Questions, HotpotQA, and held-out code tasks, multi-agent RL improves manually specified workflows after optimization, with especially large gains for smaller models and strict code all-passed metrics. These results show that UnityMAS-O can serve as a reusable substrate for converting diverse LLM-based multi-agent workflows into trainable multi-agent RL systems.

In this work, we study Cooperative Spatial Intelligence, the ability of decentralized embodied agents to coordinate effectively under dynamic environmental constraints across city-scale outdoor domains. We introduce Sentinel Challenge, a benchmark where multiple decentralized embodied agents must communicate in natural language to agree on a mutually safe and convenient meeting point within large, city-scale outdoor environments. Each agent must then navigate safely while avoiding dynamic sentinels patrolling the area, using a tool that provides coarse spatial information. To address this, we propose CoSaR (Cooperative Spatial Reasoning and Planning), a framework that bridges the high-level communication and planning abilities of foundation models with the precision of classical spatial navigation algorithms. CoSaR enables agents to exchange situational updates, reason over evolving spatial constraints, and collaboratively replan trajectories. Evaluated across 14 city-level scenes with 3-5 agents, CoSaR consistently leads to faster gathering, shorter path lengths, and improved safety. Our results demonstrate that integrating dynamic communication with spatial reasoning is essential for robust multi-agent cooperation. By formalizing this new setting and providing a scalable benchmark, we aim to build a foundation for advancing cooperative spatial intelligence in embodied multi-agent systems. Code and challenge are available at https://github.com/UMass-Embodied-AGI/Sentinel.

The effectiveness of multi-agent LLM deliberation depends not only on the agents' individual predictions, but also on how they communicate and collaborate. We study this mechanism through the lens of Friedkin-Johnsen (FJ) opinion dynamics, a tractable model for analyzing stubbornness, influence, and opinion change in multi-agent systems that captures empirically observed deliberation patterns. We show that the FJ parameters are input-dependent, turning multi-agent deliberation into a mixture of experts. This perspective implies that multi-agent systems can outperform single agents and static ensembles when routing reflects agent competence. Since competence is latent in practice, we analyze how influence is established through observable proxies: agents' self-assessed confidence, their perceived confidence, and initial alignment with other agents' views.

Navigating teams of unmanned vehicles through environments containing dynamic, adversarial Weapon Engagement Zones~(WEZs) poses a fundamental challenge to mission success: a single vehicle, however capable its onboard guidance, remains a single point of failure. This paper presents a role-differentiated multi-agent framework for collaborative threat-aware trajectory planning in which a fleet of Autonomous Collaborative Platforms~(ACPs) is assigned distinct roles primary intercept, escort, and decoy to improve team-level mission success probability while managing individual WEZ exposure. Each ACP independently employs a reactive guidance law derived from the Collision Sphere Boundary for Evader Zero-Set~(CSBEZ), which accounts for pursuer maneuverability constraints imposed by minimum turn radius, and steers the vehicle toward the safest heading that also makes progress toward its goal. Role assignment and spatial route separation induce two complementary effects: probabilistic redundancy, in which $N$ independent paths raise the team success probability and threat saturation, in which lower-priority escorts and decoys draw adversary attention and free the primary vehicle to transit uncontested.

Density functional theory (DFT) serves as the basis for computational discovery in materials science and chemistry, yet each calculation demands extensive human effort: adjusting algorithms when convergence stalls, revising plans when unexpected physics emerges, and inserting steps as intermediate results reshape the problem. Existing LLM-based agents automate only the initial planning stage, producing a full execution plan upfront and leaving all subsequent adaptation to hand-crafted rules. As a result, these workflows remain fragile, do not generalize well beyond pre-planned scenarios, and often require expert intervention when failures or unexpected intermediate results require changes to the calculation path. Here, we introduce AutoDFT, a closed-loop multi-agent framework that embeds LLM reasoning into every stage of the DFT lifecycle, where a strategic planner produces a skeletal plan of step objectives; a step planner generates numerical parameters just in time from preceding results; and a monitor-recover-reflect cycle diagnoses failures, repairs them, and revises the plan when the evidence justifies it. We demonstrate both breadth and depth: breadth on VASPBench, a purpose-built benchmark spanning 34 tasks and 9 DFT calculation types, where AutoDFT achieves 94.1% task-level success with GPT-5.2; and depth on established materials databases, where AutoDFT produces quantitatively reliable property predictions across electronic, magnetic, and energetic properties. By closing the loop between planning and execution, AutoDFT enables experimentalists without deep computational expertise to obtain reliable first-principles results.

Communication protocol design is a central challenge in large language model-based multi-agent systems. Existing single-channel approaches face an inherent communication trilemma: text-based methods are interpretable but verbose, while latent-space methods are efficient but opaque and limited to unidirectional workflows. Inspired by multi-channel communication theory, we propose HyLaT, a hybrid latent-text communication protocol that transmits elaborate cognitive signals through a latent channel for efficiency, while expressing concise critical signals in natural language to preserve interpretability and precision. We introduce a two-stage training framework combining single-agent hybrid generation learning and multi-agent interactive co-training, enabling agents to generate and interpret hybrid messages across multiple rounds of interaction. Experiments demonstrate that HyLaT reduces communication overhead significantly while maintaining competitive task performance, with strong generalization and robustness across diverse settings.

While Large Language Model-based Multi-Agent Systems (LLM-MAS) demonstrate remarkable capabilities in solving complex tasks by orchestrating specialized agents and external tools, the implicit trust in tool outputs creates a critical attack surface. Existing tool attacks are limited by domain specificity or fixed and static templates. To address these challenges, we propose Evo-Attacker, which formulates the tool attack as a self-evolving, memory-augmented reinforcement learning process. Evo-Attacker constructs a dynamic attack memory and employs deliberative reasoning to retrieve adversarial patterns and strategize modifying interventions at critical moments. Furthermore, we introduce Attack-Flow GRPO to optimize intermediate reasoning steps via terminal outcomes, addressing the long-horizon credit assignment challenge. Comprehensive experiments demonstrate that Evo-Attacker consistently outperforms baselines, highlighting its generalization and evolutionary capabilities and the urgent need for defensive tool safeguards.

Recursive Multi-Agent Trading System (RMATS) integrates four specialized agents -- Sentiment, Report, Analysis, and Risk -- coordinated through a recursive Manager Agent with iterative feedback loops. Experimental evaluation over a 561-trading-day period (January 2023 to March 2025) across a 24-asset multi-class universe demonstrates that RMATS achieves a maximum drawdown of 9.62%, lower than MVO (15.49%) and FinBERT Sentiment (15.28%), and exhibits the lowest event-period drawdown in 3 of 5 geopolitical stress scenarios tested. While RMATS underperforms return-maximizing baselines in a sustained bull market environment, ablation studies confirm the individual contribution of each agent component to downside protection. These results position RMATS as a risk-control-oriented architecture suitable for institutions prioritizing capital preservation under geopolitical uncertainty.

AI agents are increasingly used to solve complex, multi-step tasks, but existing multi-agent frameworks remain brittle as workflows grow in scale and depth. Small errors at intermediate stages can propagate through agent interactions, while insufficient grounding and weak verification mechanisms further limit reliability. We present Meta-Agent, a two-phase framework that automatically constructs and executes specialized multi-agent systems from natural-language task descriptions. In the construction phase, a task planner decomposes a problem into a directed acyclic graph of agent specifications with explicit input/output contracts and verification criteria. A web search module grounds each specification with external evidence, and a code generation module produces system prompts and tool configurations. A construction-time verification stage then validates generated artifacts and triggers targeted regeneration when failures are detected. In the execution phase, a coordinator dispatches subtasks across the agent graph while execution-time verification gates intermediate outputs. We further introduce a three-level error attribution mechanism that distinguishes local, upstream, and structural failures, enabling targeted recovery strategies ranging from localized retries to partial re-execution and re-decomposition. We evaluate Meta-Agent across coding, contextual learning, and open-ended reasoning tasks. Experiments against strong multi-agent baselines and ablation studies demonstrate consistent improvements in task success rate, error recovery, and workflow stability. The results highlight the importance of tightly integrating planning, grounding, and verification for building reliable multi-agent systems.

Industrial asset operations and maintenance question answering is inherently multi-turn, iterative, and highly dependent on external tool invocation. However, the conventional plan-execute single-agent architecture exhibits clear limitations in maintaining cross-turn context, and reusing intermediate results. In this paper, we present a multi-turn dialog system designed for industrial scenarios based on a supervisor-specialist multi-agent architecture. To alleviate tool invocation bottlenecks, the system incorporates structured artifact reuse, dynamic replanning, and parallel tool execution. Evaluation results show that our system achieves better response quality compared with the baseline, with planning effectiveness increasing by 54.5% and task completion improving by 37.8%. System profiling further shows that cross-turn artifact reuse effectively reduces redundant tool invocation, decreasing the tool-time share from 47.3% to 26.3% and making turns 2-5 approximately 4.2x faster than the first turn.

Manufacturing has passed through four widely recognized paradigms - mechanization, electrification, programmable automation, and Smart Manufacturing - each defined by the kind of work it shifted from humans to machines. In every case, one layer of industrial work remained fundamentally human: the coordinative cognition of production, comprising the interpretive, allocative, diagnostic, negotiative, and governance work exercised by engineers, planners, and operational managers. We argue that a fifth transition is now underway in which this layer, rather than the physical or routine-cognitive layers below it, is what foundation-model-based autonomous agents primarily redistribute. We name this paradigm Agent Manufacturing and define it operationally: a manufacturing system is an instance of Agent Manufacturing when its principal coordination mechanism is reasoning performed by foundation-model agents that can interpret open-ended goals, plan over long horizons, invoke tools and machines, and negotiate with other agents and humans. This is a narrower and more falsifiable definition than the existing literature on cognitive manufacturing or Industry 5.0 provides, and it distinguishes the paradigm sharply from classical multi-agent manufacturing systems, which were autonomous only within closed protocol spaces.

Operating LLMs as coordinated multi-agent research systems over multi-hour runs surfaces failure modes that single-shot evaluation cannot: upstream providers throttle without warning, sub-agents drift the task to fit accessible tools, narrate machinery instead of using it, open revision iterations with self-apology, or treat upstream context as executable directives. We present PRIMA, whose primary contributions are three operational patterns for surviving these failure modes: (1) a resilience-and-recovery layer that detects upstream rate-limit signals, persists a typed pause record to disk, and resumes long-running runs without re-executing converged work even across process restarts; (2) a sub-agent operating discipline encoding task-fidelity, tool-use, revision, and inter-step context-boundary norms as a structural prompt layer; (3) a multi-phase application pattern for structured engineering deliverables pairing orthogonal draft steps with an explicit cross-document harmonization pass before final synthesis. These sit atop a foundational protocol: a research-program specification language with explicit convergence criteria, a dual-metric scoring engine (LLM-judged rubric plus sandboxed code), an outer meta-optimization loop, event-driven persistence, hook-based middleware, context compaction, and a multi-provider LLM abstraction. Agent identities derive from prime powers, giving collision-free identifiers and trivially-verifiable cluster membership without a central registry. Theoretical guarantees include $O(k)$ verification, $O(V+E)$ DAG validation, and identity collision freedom by the Fundamental Theorem of Arithmetic. A Graph Isomorphism case study grounds the architectural claims in a generated artifact: a six-step protocol that produced a research paper proposing a new canonical-form algorithm with three theorems and five conjectures.

[Abridged] - Spectral Retrieval is a plug-in re-ranking stage that interpolates between per-token MaxSim and mean-pool retrieval through a multi-scale sinc convolution over token embeddings. In standard dense retrieval each document is one mean-pooled vector; when relevance localises into a short subspan, the signal averages into noise. Spectral Retrieval reuses per-token embeddings from a late-interaction index and convolves them with a normalised sinc kernel at multiple scales. At L=1 the kernel acts as the identity, recovering per-token MaxSim; as L grows it approaches a uniform filter, recovering mean pooling. The maximum cosine over positions and scales yields a score provably no less informative than either endpoint. On a controlled synthetic benchmark with 1,000 documents and planted single-position spikes, mean-pool retrieval sits at chance (Recall@10 ~ 0.02) regardless of spike strength, while Spectral Retrieval reaches Recall@10 = 1.0 once the planted cosine exceeds the corpus-level token noise floor. On LIMIT-small with a frozen all-mpnet-base-v2 encoder, Spectral Retrieval lifts Recall@10 from 0.33 to 0.90, MRR from 0.22 to 0.79, and strict Success@10 from 0.12 to 0.84, without retraining. The method fits naturally into multi-agent LLM systems, where each agent benefits from a tighter, role-specific retrieval window over a shared corpus.

Large Language Models (LLMs) are increasingly applied to software engineering (SE), yet their potential for autonomous, role-oriented collaboration remains largely underexplored. Understanding how multiple LLM-based agents coordinate, maintain role alignment, and converge on solutions is critical for SE, as naively allowing agents to interact does not reliably lead to correct or stable outcomes. Recent empirical studies show that unstructured or poorly understood interaction dynamics can result in error propagation, premature consensus on incorrect solutions, or prolonged disagreement that prevents convergence, even when correct partial solutions are present early in the interaction. As an initial step towards addressing this underexplored area, we undertake a systematic analysis of conversations between two agents, a Designer and a Programmer across 12 model combinations from 7 open-source LLMs (Gemma 2, Gemma 3, LLaMA 3.2, LLaMA 3.3, DeepSeek-R1, MiniCPM, and Qwen3). Our systematic approach reveals three key dimensions of multi-agent interaction: efficiency (the speed and stability of convergence), consistency (the degree of role alignment visualized by BLEU and ROUGE), and effectiveness (the extent of compilation success and error resolution). Results show that the DeepSeek-R1:DeepSeek-R1 pair was unique in converging to the correct solution from the very first iteration and sustaining it consistently to the final iteration, while LLaMA 3.2:LLaMA 3.2 and Qwen3:Qwen3 demonstrated strong Designer:Programmer role alignment despite of diverging from the correct solution. The other pairs deviated from the task, never to converge to a result. These findings advance understanding of agentic programming and highlight the need for further research on understanding and calibrating convergence and stop conditions essential for future autonomous SE.

LLM-based multi-agent systems can fail even when planned actions are executed correctly because agents may misjudge their knowledge when evaluating plan feasibility, a phenomenon we term epistemic miscalibration in planning. Unlike execution errors, epistemic miscalibration is latent during planning, as generated plans can remain self-consistent and executable without observable errors; the miscalibration is also dynamic, as new information can alter feasibility assessments, potentially obscuring past miscalibration signals and causing them to recur over time. To address this, we propose the Epistemic Planning Calibration Agentic Workflow (EPC-AW), which assesses whether plans remain supported under varying information conditions rather than directly verifying feasibility. EPC-AW employs Information-consistency-based Plan Selection, selecting plans whose evaluations are stable across agents, together with Consistency-guided Epistemic State Refinement to adapt calibration over time by leveraging past discrepancies to guide future planning. Experiments show that EPC-AW improves system-level success by an average of 9.75%.

With the increasing complexity of collaboration among various social entities and user demands, the factors affecting the stable development of the data service market are also growing. These factors include the widespread dissemination of information enhancing subjective consciousness, the continuous improvement in intelligence, and the complexification of structural relationships. To achieve effective governance and regulation of the data service market, it is crucial to conduct simulation experiments before making regulatory decisions. However, current research and analysis of the data service market primarily focus on data-level performance, proving inadequate when it comes to measurement and analysis of multiple heterogeneous entities and the integration of various social elements within the data service market. Based on this, this paper innovatively proposes a data service market measurement and network analysis method based on heterogeneous multi-agent modeling. By introducing the service ecosystem theory, we clarify the participants and external factors of the data service market and conduct utility measurements for three-level entities based on value creation. Furthermore, an analytical methodology is devised to precisely assess the influence of heterogeneous networks on utility. Finally, the paper verifies the effectiveness of the proposed method through the analysis of experimental results.

Topology optimization is a widely used design method that produces optimized material distributions for prescribed objectives and constraints through well-established numerical algorithms. Throughout the workflow, engineers make a series of decisions ranging from setting and adjusting numerical parameters to assessing whether the converged design meets considerations beyond those explicitly included in the optimization problem, such as physical feasibility. These decisions, which draw on domain expertise, interfere with the autonomous design process. To address this difficulty, this study presents TopOptAgents, a multi-agent system for automating not only the design process but also decision-making during the key stages of the topology optimization process. TopOptAgents consists of six LLM-based agents collaborating through iterative self-refinement cycles spanning problem formulation, validation, code generation and execution, and quality assessment of the optimized structure. This process enables error correction and progressive improvement of both the optimization setup and resulting design. The framework is demonstrated on optimization problems selected to cover a range of settings that differ in their literature coverage and numerical characteristics The benefits of iterative self-refinement are found to be particularly pronounced for problem classes where the pretrained language model has limited prior exposure, such as formulations whose literature and open-source implementations are comparatively sparse. In such cases, the proposed framework reliably produces converged designs where a single state-of-the-art LLM struggles, suggesting that self-refinement broadens the range of topology optimization problems that LLM-based automation can reliably address.

Gardening is critical to support well-being, cultural continuity, and food autonomy, yet existing digital tools often provide generic advice that overlooks gardeners' skills, local ecologies, seasons, and cultural contexts. We introduce CultivAgents, a relationship-centered multi-agent system for personalized, socio-culturally grounded gardening support. Grounded in ethics of care, CultivAgents coordinates multiple specialized agents: an Experience Agent that adapts guidance to users' skill levels, an Environmental Agent that grounds advice in local and seasonal conditions, and an Ethnobotanical Agent that connects plants to cultural knowledge and histories. We evaluated CultivAgents through a three-phase mixed-methods study with domain experts (n=3), HCI researchers (n=7), and community gardeners (n=5), analyzing expert feedback, pre/post surveys, and participatory design activities. Results suggest that CultivAgents helped gardeners translate interest into situated action: community gardeners reported increased confidence (3.00 to 3.60), motivation (4.00 to 4.40), and trust in acting on AI advice (3.20 to 4.00). Participants valued hyperlocal ecological guidance and complementary agent perspectives, while also identifying limits in cultural specificity, ecological grounding, and agent coordination. The work advances relationship-centered AI, offering design implications for multi-agent systems that support food sovereignty, community resilience, and cultural preservation.

In orchestrated multi-agent systems, humans often struggle to manage plans due to their complexity and limited transparency. Existing approaches rely on outcome-level supervision, where users verify only final outputs without visibility into intermediate reasoning. We formalize a design space for human-LLM co-planning interactions along three axes: mode (semantic vs. structural), scope (global vs. targeted), and level (low vs. high-level edits). We realize it in AMBIPOM, a prototype supporting process-level supervision through both semantic and structural interactions. Through a user study, we characterize how users navigate this space, revealing hybrid workflows and effort-control-risk trade-offs; through a controlled benchmark, we analyze how LLMs revise plans under varying scope and revision strategies. Our findings yield design insights for more transparent, controllable, and effective human-AI co-planning. We release code and data at https://github.com/megagonlabs/ambipom.

Large language model (LLM)-based multi-agent systems increasingly rely on intermediate communication to coordinate complex tasks. While most existing systems communicate through natural language, recent work shows that latent communication, particularly through transformer key-value (KV) caches, can improve efficiency and preserve richer task-relevant information. However, KV caches also encode contextual inputs, intermediate reasoning states, and agent-specific information, creating an opaque channel through which sensitive content may propagate across agents without explicit textual disclosure. To address this, we introduce \textbf{LCGuard} (Latent Communication Guard), a framework for safe KV-based latent communication in multi-agent LLM systems. LCGuard treats shared KV caches as latent working memory and learns representation-level transformations before cache artifacts are transmitted across agents. We formalize representation-level sensitive information leakage operationally through reconstruction: a shared cache artifact is unsafe if an adversarial decoder can recover agent-specific sensitive inputs from it. This leads to an adversarial training formulation in which the adversary learns to reconstruct sensitive inputs, while LCGuard learns transformations that preserve task-relevant semantics and reduce reconstructable information. Empirical evaluations across multiple model families and multi-agent benchmarks show that LCGuard consistently reduces reconstruction-based leakage and attack success rates while maintaining competitive task performance compared to standard KV-sharing baselines.

Self-evolving multi-agent systems (MAS) have emerged as a promising route to LLM agents that continually improve from experience, with persistent memory at their foundation. However, existing designs almost exclusively adopt a centralized repository shared across agents, incurring communication and coordination overhead, raising privacy concerns, and collapsing agent diversity. We propose DecentMem, a decentralized memory framework in which each agent maintains its own dual-pool memory -- an exploitation pool of consolidated past trajectories and an exploration pool of LLM-generated candidates for unseen contexts. The two pools are reweighted online based on stage-wise feedback from an LLM-as-a-judge. Theoretically, we prove that this design guarantees global reachability of the solution space and achieves $O(\log T)$ cumulative regret, matching the stochastic bandit lower bound up to constants. In practice, across three MAS frameworks (AutoGen, DyLAN, AgentNet), three Qwen3 backbones (4B/8B/14B), two Gemma4 backbones (E2B/E4B) and five benchmarks spanning math, code, QA, and embodied tasks, DecentMem improves average accuracy by up to 23.8% over the strongest centralized memory baseline and by up to 52.5% over the no-memory baseline, while reducing token usage by up to 49%.

Existing approaches for digital short-drama production typically rely on one-shot LLM generated scripts and loosely coupled pipelines, which fail to satisfy three key requirements of short-drama generation: (1) narrative pacing, resulting in weak hooks, insufficient escalation, and unattractive endings; (2) spatial consistency, leading to drifting scene layouts and inconsistent character positions across clips; and (3) production-level quality control, requiring extensive manual review and correction across script and visual stages. We present One Sentence, One Drama, a hierarchical multi-agent framework that transforms a user's single-sentence idea into a fully produced short drama through structured intermediate modules and iterative refinement. Our approach is built upon three key components: (1) a multi-agent debate-based story generation module that enforces short-drama pacing and narrative coherence; (2) a 3D-grounded first-frame generation mechanism that establishes a shared spatial reference for consistent character positioning and scene layout across clips; and (3) multi-stage reviewer loops that perform comprehensive error detection and targeted revision across script, visual, and video generation stages. We also introduce scene-level BGM matching and scene transition planning to improve the audience's immersive experience. To systematically evaluate this task, we introduce Short-Drama-Bench, a benchmark that extends standard video quality metrics with short-drama-specific criteria. Experimental results demonstrate that our method significantly outperforms existing pipelines in narrative quality, cross-clip consistency, and overall viewing experience.

The proliferation of emerging applications, such as autonomous driving and immersive experiences, demands cellular networks that are not only faster, but fundamentally more resilient and autonomous. This paper presents a BlueSky vision on how Artificial Intelligence will be natively integrated into 6G, shifting the paradigm from \underline{Network for AI} to \underline{AI for Network}. We envision that, unlike 5G's reliance on scattered, ad-hoc models each trained for a single task, native AI in the 6G era will be anchored by a foundation model and and orchestrated via collaborative multi-agent systems, framing network management as a unified, multi-modal, multi-task optimization problem. Built on this vision, we outline two transformative directions. The first focuses on developing a 6G foundation model as a unified backbone, with task-specific knowledge distilled into compact models suited for diverse edge deployments. The second advances multi-agent systems designed to autonomously diagnose, maintain, and recover networks with minimal human intervention. These directions chart a roadmap for 6G to evolve into an intelligent, self-sustaining communication infrastructure.

Deploying neural networks on microcontroller units (MCUs) is critical for edge intelligence but remains challenging due to tight memory, storage, and computation constraints. Existing approaches, such as model compression and hardware-aware neural architecture search (HW-NAS), often depend on proxy metrics, incur high search cost, and do not fully bridge the gap between architecture design and verified deployment. This paper presents AutoMCU, a feasibility-first large language model (LLM)-based multi-agent system for automated neural network customization under MCU constraints. Given natural-language task requirements and hardware specifications, AutoMCU iteratively generates structured architecture candidates, filters infeasible designs through vendor toolchain feedback before training, evaluates feasible models under a controlled protocol, and verifies deployability through backend-grounded deployment analysis. AutoMCU includes two key mechanisms: 1) hardware-in-the-loop architecture generation for early elimination of undeployable candidates under RAM and Flash constraints, and 2) state-isolated multi-agent scheduling for stable coordination of proposal, training, evaluation, and deployment stages. Experiments on CIFAR-10 and CIFAR-100 under strict MCU constraints show that AutoMCU achieves competitive accuracy while reducing customization time to about 1--2 hours, compared with hundreds of GPU hours for representative MCU-oriented HW-NAS baselines. Comparisons with ColabNAS and the LLM-based NAS method GENIUS on NAS-Bench-201 further demonstrate the effectiveness and stability of AutoMCU. Real-device deployments on multiple STM32 microcontrollers validate its practical applicability to MCU-scale edge intelligence.

Although Vehicle Routing Problems (VRP) are essential to many real-world systems, they remain computationally intractable at scale due to their combinatorial complexity. Traditional heuristics rely on handcrafted rules for local improvements and occasional \textit{jumps} to escape local minima, but often struggle to generalize across diverse instances. We introduce \textbf{COAgents}, a cooperative multi-agent framework that models the search process as a graph: nodes represent solutions, and edges correspond to either local refinements or large perturbations for diversification (i.e., jumps). A \textit{Partial Search Graph} (PSG) is dynamically constructed during search, enabling COAgents to train a Node Selection Agent and a Move Selection Agent to guide intensification, and a Jump Agent to trigger well-timed explorations of new regions. Unlike end-to-end learning approaches, COAgents cleanly separates problem-agnostic search control from compact domain-specific encoding, facilitating adaptability across tasks. Extensive experiments on the CVRP and VRPTW benchmarks show that COAgents remains competitive with several learn-to-search baselines on CVRP and sets a new state of the art among learning-based methods on the more challenging VRPTW instances, reducing the gap to the best-known solutions by 14% at $N\!=\!100$ and 44% at $N\!=\!50$ relative to the strongest neural solver (POMO), and by 21% and 40% respectively relative to ALNS. Code is available at https://github.com/mahdims/COAgents.

Realizing Level 4/5 Autonomous Networks (AN) demands a shift from static automation to agent-native intelligence. Current operations, reliant on rigid scripts, lack the cognitive agency to handle off-nominal conditions. To address this, this letter proposes a hierarchical multi-agent reference architecture enabling high-level autonomy. The framework features a Dual-Driven Orchestrator that coordinates specialized Executive Agents, supported by a shared Public Memory for unified domain knowledge. A key innovation is the integration of agent self-awareness, which empowers the system to harmonize deliberative strategic governance with reflexive fault recovery. We instantiate and validate this architecture within a 5G Core environment. Case studies demonstrate that the system sustains critical throughput under congestion and reduces Mean Time to Repair (MTTR) by 86%, confirming its efficacy in unifying strategic planning with operational resilience.

Recent advances in multi-agent systems have shown great potential for solving complex tasks. However, when multiple agents edit a shared codebase concurrently, their changes can silently conflict and inconsistent views lead to integration failures. Existing multi-agent systems address this through workspace isolation (e.g., one git worktree per agent), but this defers conflict resolution to a post-hoc merge step where recovery is expensive. In this paper, we propose STORM, i.e., STate-ORiented Management for multi-agent collaboration. Specifically, STORM manages agent states by mediating their interactions with the shared workspace, ensuring that each agent operates on a consistent view of the codebase and that conflicting edits are detected and resolved at write time. We evaluate STORM on Commit0 and PaperBench across multiple LLMs. STORM outperforms the git-worktree-based multi-agent baseline by +18.7 on Commit0-Lite and +1.4 on PaperBench, while achieving comparable or better cost efficiency. Combined with single-agent runs, STORM reaches highest scores of 87.6 and 78.2 on the two benchmarks respectively, suggesting that explicit state management is a more effective foundation for multi-agent collaboration than workspace isolation. STORM can also be plugged into any multi-agent system seamlessly.

Large Language Models (LLMs) have enabled collaborative Multi-Agent (MA) systems, where interacting agents improve performance through diverse reasoning and iterative refinement. However, these systems remain vulnerable to error propagation, where early-stage information degrades downstream reasoning. To address this, we conduct a systematic analysis of inter-agent communication to identify which information drives MA performance. We find that the absence of reasoning and verification in inter-agent communication significantly degrades performance. Based on these insights, we propose Category-Aware Recovery Augmentation (technique), which enforces the presence of critical information during communication. recovers up to 86.2% of failed cases. Our results highlight the key role of information quality in effective MA collaboration. Our code is available at https://anonymous.4open.science/r/cara_mas

Large Language Model (LLM) agents are increasingly applied to engineering design tasks, yet existing evaluation frameworks do not adequately address multi-agent systems that combine simulation, retrieval, and manufacturing preparation. We introduce a benchmark suite with three evaluation dimensions: (1) a workflow benchmark with seven prompt styles targeting distinct cognitive demands-including direct tool use, semantic disambiguation, conditional branching, and working-memory tasks; (2) a Retrieval-Augmented Generation (RAG) benchmark with gated scoring isolating retrieval contributions to parameter selection; and (3) an High Performance Computing (HPC) benchmark evaluating end-to-end ML training orchestration on a SLURM cluster. Alongside the benchmark we present EngiAI, a Multi-Agent System (MAS) reference implementation built on LangGraph that operationalizes the benchmark by coordinating seven specialized agents through a supervisor architecture, unifying topology optimization, document retrieval, HPC job orchestration, and 3D printer control. Across four LLM backends and two EngiBench problems, proprietary models achieve 96-97% average task completion on Beams2D, while open-source 4B-parameter models reach 55-78%, with clear generational improvement. Conditional branching proves most challenging, with task completion dropping to 20-53% for the conditional style on Photonics2D. RAG gating confirms near-perfect retrieval-augmented scores (about 1.0) versus near-zero without retrieval, validating the evaluation design. On HPC orchestration, one model completes all pipeline steps in 100% of runs while another drops to 50%, revealing that multi-step instruction following degrades over long-running workflows.

LLM-based multi-agent systems (MAS) have demonstrated strong reasoning and decision-making capabilities that consistently surpass those of single LLM agents. However, their performance often suffers from naive aggregation mechanisms that assume uniformly cooperative interactions. Upon close inspection, we observe that existing graph-based MAS frameworks (1) propagate errors when conflicting signals arise without control, and (2) lack explicit modeling of conflicting inter-agent relations as well as structural awareness, failing to identify reliable interaction patterns. To bridge this gap, we introduce SIGMA, a novel SIgned Graph-informed Multi-Agent reasoning framework that explicitly captures trust, conflict, and neutral relations among agents via a signed relational graph. Specifically, given a query, SIGMA first selects a set of relevant and diverse agents, then constructs a structured signed interaction graph with confidence-weighted edges. Reasoning proceeds through conflict-aware signed message passing, which reinforces information from trustworthy agents while suppressing conflicting signals, and terminates with a structure- and conflict-aware weighted aggregation to yield globally consistent and conflict-resilient predictions. Extensive experiments on six benchmark datasets, across multiple LLM backbones and diverse multi-agent configurations, demonstrate that SIGMA consistently outperforms state-of-the-art baselines, achieving notable gains in both accuracy and conflict-resilient performance.

Cascade attacks in LLM multi-agent systems (MAS) arise when adversarial influence propagates across agents and leads to escalated system-level failures through complex agent interactions. Detecting such cascades is challenging, as their signals are distributed, tightly coupled across interaction channels, and often appear plausibly benign locally but may unfold quickly either within a single turn or gradually across multiple turns. Existing defenses, being largely local and text-centric, fail to capture such cross-channel, temporally coordinated dynamics of cascade propagation. Therefore, we propose CASPIAN, the first framework that provides a unified, cross-channel causal analysis of cascade behavior in LLM-MAS through online monitoring of dynamic influence propagation across agents. CASPIAN models multi-agent interactions using a unified, dynamic causal influence matrix across channels, estimated efficiently via a late-interaction conditional transfer entropy (LI-CTE) formulation, thereby enabling the detection of cascade onset from emergent system-level structure rather than isolated anomalies. It further performs online causal attribution, identifying the origin, bridge, and amplifier agents driving the cascade and reconstructing its principal propagation pathways, capabilities not supported by existing methods. Across diverse multi-agent frameworks and benchmarks, CASPIAN consistently outperforms semantic guardrails, LLM-based judges, and graph-based anomaly detectors in both detection accuracy and early cascade identification while operating with sub-1% relative overhead latency. These results demonstrate that unified cross-channel causal modeling is essential for reliably detecting and understanding cascade failures in LLM multi-agent systems.

The rapid advancement of autonomous driving (AD) technologies has outpaced the development of robust safety evaluation methods. Conventional testing relies on exposing AD systems to vast numbers of real-world traffic scenes -- a brute-force approach that is prohibitively expensive and statistically ineffective at capturing the rare, safety-critical edge cases essential for validating real-world robustness. To address this fundamental limitation, we introduce STRELGen, a scalable framework for the targeted generation of safety-critical driving scenarios. STRELGen synergistically combines a multi-agent trajectory-generation diffusion model (DM) with Spatio-Temporal Logic (STREL) specifications that encode complex safety and realism properties through a highly interpretable formalism. Crucially, monitoring satisfaction levels of these specifications is differentiable, enabling gradient-based search. At inference time, we optimize directly over the DM latent space to maximize STREL formula satisfaction. The result is efficient generation of highly plausible yet safety-critical multi-agent scenarios that lie within the learned data distribution. STRELGen thus provides a flexible, interpretable, and powerful tool for stress-testing autonomous driving systems, moving beyond the limitations of brute-force data collection.

Recent large language models (LLMs) have demonstrated strong capabilities in understanding and generating code, from competitive programming to repository-level software engineering. In emerging agentic systems, code is no longer only a target output. It increasingly serves as an operational substrate for agent reasoning, acting, environment modeling, and execution-based verification. We frame this shift through the lens of agent harnesses and introduce code as agent harness: a unified view that centers code as the basis for agent infrastructure. To systematically study this perspective, we organize the survey around three connected layers. First, we study the harness interface, where code connects agents to reasoning, action, and environment modeling. Second, we examine harness mechanisms: planning, memory, and tool use for long-horizon execution, together with feedback-driven control and optimization that make harness reliable and adaptive. Third, we discuss scaling the harness from single-agent systems to multi-agent settings, where shared code artifacts support multi-agent coordination, review, and verification. Across these layers, we summarize representative methods and practical applications of code as agent harness, spanning coding assistants, GUI/OS automation, embodied agents, scientific discovery, personalization and recommendation, DevOps, and enterprise workflows. We further outline open challenges for harness engineering, including evaluation beyond final task success, verification under incomplete feedback, regression-free harness improvement, consistent shared state across multiple agents, human oversight for safety-critical actions, and extensions to multimodal environments. By centering code as the harness of agentic AI, this survey provides a unified roadmap toward executable, verifiable, and stateful AI agent systems.

This position paper argues that enforcing LLM agent safety within a single abstraction layer is not merely suboptimal but categorically insufficient for deployed LLM agents -- a structural consequence of how agent execution works, not a contingent limitation of current systems. The three dimensions that jointly constitute safe operation -- semantic intent and policy compliance, environmental validity, and dynamical feasibility -- each depend on a strictly distinct set of information that becomes available at different stages of execution. No single guardrail can certify all three. We argue that the community must respond with a contract-based architecture in which each safety dimension is enforced by an independently certified layer whose probabilistic guarantee satisfies the next layer's assumption. We sketch such an architecture and derive the compositional system-level safety bounds it admits via the chain rule of probability. Three open problems stand between this and a deployable standard: bound estimation from non-i.i.d.\ traces, graceful degradation of contracts under deployment drift, and extension to multi-agent settings -- the most important unfinished business in LLM agent runtime assurance.

Although artificial intelligence (AI) now matches or exceeds human performance across numerous cognitive tasks, creativity remains a highly contested frontier. As AI systems based on large language models (LLMs) are increasingly adopted in research and innovation, it is essential to understand and augment their creativity. Here we demonstrate that multi-agent LLM teams not only surpass single agents, but also substantially outperform human teams in creativity (Cohen's d=1.50) across 4,541 multi-agent LLM ideas and 341 human-team ideas on six diverse problem-solving tasks. This advantage is driven by novelty while maintaining comparable usefulness. To investigate the generative processes in both groups, we represent conversations as paths through semantic space using neural language model representations. Both LLM and human teams produce more creative ideas when conversations range widely rather than staying centered on a single theme (low global coherence). However, the additional patterns that predict creativity differ: LLM teams benefit from efficient exploration (high semantic spread, shorter paths), while human teams benefit from maintaining smooth conversational flow (high local coherence, frequent pivots). Additionally, we identify model choice and discussion structure as orthogonal design levers that together explain 26.8% of variance in LLM conversational dynamics, paving the way for systematic approaches to developing multi-agent systems with augmented creative capabilities.

Multi-agent systems extend large language models (LLMs) by decomposing tasks among specialized agents, but their distributed decision process creates new attack surfaces. We identify semantic hijacking, an attack in which harmful requests are concealed within domain-specific narratives and propagated to a Manager through Worker reports, without any syntactic injection primitives. Across 42,000 adversarial trials over 12 Manager models and 7 Worker configurations, we uncover a capability paradox: as Worker capability increases, the mean system-level Attack Success Rate (ASR) increases from 18.4% to 63.9%, peaking at 94.4%. To explain this effect, we conduct multi-level mediation analysis on two independent datasets (47,807 interactions). This analysis shows that this paradox is driven by linguistic certainty: stronger Workers are more likely to interpret adversarial narratives as legitimate, convey their conclusions assertively, and thereby lead Managers to treat such confident endorsements as justification to execute. In our larger Worker-Only setting ($n_W$=14), certainty mediates 74% of the effect, with 95% confidence intervals (CI) excluding zero under both Monte Carlo and cluster bootstrap; the smaller Full-MAS setting ($n_W$ =6) shows a directionally consistent indirect effect. Worker-side safety prompting does not reliably mitigate this failure. Building on the mediation finding, we propose heterogeneous ensemble verification, which pairs Workers of asymmetric domain competence so their complementary vulnerabilities break the certainty-to-execution chain, reducing ASR from 52.8% to 2.0% with negligible benign-task impact. Our results show that upgrading components to stronger models can actively degrade system security, and that effective defenses require exploiting--rather than eliminating--capability asymmetries between agents.

Large language model-driven multi-agent systems (LLM-MAS) excel at complex tasks, yet unreliable agents remain a key bottleneck to system-level reliability. Automatic failure attribution is therefore critical, but existing approaches, such as direct prediction of agent-error pairs and agent-first failure attribution, rely on local logs of agents and miss global failures that only manifest over full interaction trajectories, such as cross-step inconsistencies and inter-agent coordination errors. Moreover, directly predicting failures induces a large combinatorial search space, hindering fine-grained attribution. To address these challenges, we propose VerifyMAS, a hypothesis verification framework for agent failure attribution. Instead of directly predicting faulty agents and error types, VerifyMAS formulates and verifies failure hypotheses against full trajectories. This verification-based approach decomposes attribution into trajectory-level error validation and fine-grained agent localization, providing an error-first attribution approach that captures global failure patterns while substantially reducing the search space. We further introduce a hypothesis-based data construction strategy grounded in a structured error taxonomy and fine-tune a specialized LLM verifier model for trajectory-level failure verification and agent attribution. Experiments on Aegis-Bench and Who&When show that VerifyMAS consistently improves diverse backbone models, including open-source Qwen and API-based GPT models, outperforming prior methods without sacrificing inference efficiency for long multi-agent trajectories.

The rapid evolution of Large Language Model (LLM) agents has produced diverse interaction paradigms, yet few production systems integrate multiple paradigms within a unified architecture. This paper presents a systematic analysis of three principal agent interaction paradigms, including Multi-Agent Orchestration (Generator-Evaluator), ReAct Tool-Use Loops, and Memory-Augmented Interaction, as implemented in buddyMe, an open-source multi-model agent programming framework. We formalize a five-stage processing pipeline: Requirement Pre-Review -> Task Decomposition -> ReAct Execution -> Real-Execution Verification -> Adversarial Evaluation Discussion, and establish a six-dimensional evaluation schema with weighted scoring. Through four empirical case studies drawn from real-world deployment logs covering museum guide generation, scheduled weather tasks, and comprehensive tour planning, we draw three key conclusions. First, Generator-Evaluator pre-review detects requirement omissions in 20 percent of complex tasks, with 80 percent tasks passing initial inspection. Second, the ReAct loop ensures stable subtask execution but leads to around 30 percent redundant tool invocations. Third, adversarial Evaluator-Defender discussions reach consensus within 2-3 rounds for nearly 70 percent of scenarios, functioning mainly for content refinement rather than logical reversal. We additionally provide three Mermaid-based architectural diagrams and conduct cross-paradigm comparisons with CrewAI, AutoGen, LangGraph, MemGPT and A-Mem across six system dimensions. The research outcomes offer practical design guidelines for constructing stable and reliable multi-paradigm agent systems.

Multi-agent language systems are often built as hand-designed workflows, where agents are assigned semantic roles and communication protocols are specified in advance. We propose NeuroMAS, a method that first treats a multi-agent language system as a trainable and scalable neural-network-like architecture with LLM agents as nodes and intermediate textual signals as edges. In NeuroMAS, agent nodes are role-free but structure-aware: the topology only determines how information can flow in general, while reinforcement learning training determines how nodes communicate, specialize, and coordinate. This formulation shifts multi-agent design from workflow engineering toward architecture design, where depth, width, connectivity, and growth protocol become scalable sources of capability. Further, we provide a theoretical perspective showing why such modular textual computation is more parameter-efficient when tasks admit hierarchical decompositions. Experiments show that NeuroMAS improves significantly over both inference-time and trained multi-agent baselines. We further find that organizational scaling is path-dependent: larger systems can be challenging to train from scratch, but become feasible when grown progressively from smaller trained systems. These results suggest that learned neural multi-agent systems are a promising scaling axis for LLMs.

Multi-agent pathfinding (MAPF) under one-shot planning is a core component of warehouse automation, yet classical formulations typically assume four-connected 2D grids with unit-time moves in four directions. To fill reality gaps while still being trackable with discrete combinatorial search, this work proposes a more practical counterpart tailored to differential-drive AGVs. We term this multi-agent warehouse pathfinding (MAWPF), featured with four constraints: (i) agent actions are restricted to straight motion and in-place rotation; (ii) rotations require multi-step costs; (iii) acceleration and deceleration are considered, and; (iv) follower collisions are prohibited to prevent rear-end crashes. To solve MAWPF efficiently, we adapt representative suboptimal MAPF algorithms-PP, LNS2, PIBT, and LaCAM-and conduct comprehensive benchmarking. Our experiments reveal that PP and LNS2 struggle to solve instances with many agents, while PIBT-based approaches achieve preferable scalability with increased solution cost. We believe that these constitute an important step toward adapting classical gridworld MAPF to operational warehouse setups.

Multi-agent systems can solve complex tasks through collaboration between multiple Large Language Model agents. Existing collaboration frameworks typically operate in either a parallel or a sequential mode. In the parallel mode, agents respond independently to queries followed by aggregation of responses. In contrast, sequential systems allow agents to communicate via a directed topology and refine one another step by step. However, both modes are inadequate for achieving the desired objectives of minimizing communication and latency while simultaneously maximizing the accuracy of the final response. In this work, we introduce a hybrid paradigm called Nexa, a trainable response-conditioned policy that bridges the gap between the two modes. Nexa begins with a parallel execution stage, embeds the resulting responses into a shared semantic space, and then predicts a sparse directed acyclic communication graph. If the graph is empty, the system remains purely parallel; if it is non-empty, the system performs one sequential message propagation. The policy is a lightweight transformer model, and the method avoids the need for external LLM judges or reward models, as well as hand-crafted test-time topology search. We formalize this hybrid execution problem, show that the resulting graph is acyclic by construction, and that the framework strictly subsumes pure parallel execution, and present a training procedure based on policy-gradient optimization. Results demonstrate that the response-conditioned policy learned by Nexa under one setting can be reused when the number of agents, the task, or the underlying agent changes, thus emphasizing the generalizability of the learned communication policy.

Autonomous multi-agent systems based on large language models (LLMs) have demonstrated remarkable abilities in independently solving complex tasks in a wide breadth of application domains. However, these systems hit critical reasoning, coordination, and computational scaling bottlenecks as the size and complexity of their tasks grow. These limitations hinder multi-agent systems from achieving high-throughput processing for highly parallelizable tasks, despite the availability of parallel computing and reasoning primitives in the underlying LLMs. We introduce the Agent-Parallel Workload Architecture (APWA), a distributed multi-agent system architecture designed for the efficient processing of heavily parallelizable agentic workloads. APWA facilitates parallel execution by decomposing workflows into non-interfering subproblems that can be processed using independent resources without cross-communication. It supports heterogeneous data and parallel processing patterns, and it accommodates tasks from a wide breadth of domains. In our evaluation, we demonstrate that APWA can dynamically decompose complex queries into parallelizable workflows and scales on larger tasks in settings where prior systems fail completely.

When multi-agent LLM systems produce bad answers, not all failures are equal: some answers are grounded in the right material but incomplete, while others are simply ungrounded and should be stopped. Current retry strategies treat both cases identically (try again and hope for the best), leaving human supervisors unable to tell whether a retry was warranted or whether the system should have halted instead. We introduce the Argent Signaling Protocol (ASP), a compact machine-readable header that accompanies every AI-generated response with structured quality signals: certainty (@C), grounding (@G), stochasticity (@S), and an assumption index that classifies the evidentiary basis of each claim. These signals enable a controller to distinguish repairable failures from containment failures and route each case differently. We evaluate ASP in two modes. In standalone mode, a 27-question document-grounded QA benchmark over the Array BioPharma/Ono license agreement compares baseline prompts against ASP-instrumented controller actions across three local GGUF models. On Qwen~(0.8B), ASP improves pass rate from 11.1% to 33.3% and mean term coverage from 36.7% to 65.4%; on Dobby~(8B), ASP produces 4 fail-to-pass recoveries, raising pass rate from 33.3% to 44.4%; on SmolLM3~(3B), ASP alternates between repair and containment per question. Aggregate improvement is meaningful (12/81 to 21/81 passes). In multi-agent mode, an ASP sidecar sits between a retrieval agent and a downstream decision agent; the sidecar blocks 100% of ungrounded upstream outputs from reaching the downstream agent (24/27 blocked, 0 ungrounded propagations).

LLM-based autonomous agents have demonstrated strong capabilities in reasoning, planning, and tool use, yet remain limited when tasks require sustained coordination across roles, tools, and environments. Multi-agent systems address this through structured collaboration among specialized agents, but tighter coordination also amplifies a less explored risk: errors can propagate across agents and interaction rounds, producing failures that are difficult to diagnose and rarely translate into structural self-improvement. Existing surveys cover individual agent capabilities, multi-agent collaboration, or agent self-evolution separately, leaving the causal dependencies among them unexamined. This survey provides a unified review organized around four causally linked stages, which we term the LIFE progression: Lay the capability foundation, Integrate agents through collaboration, Find faults through attribution, and Evolve through autonomous self-improvement. For each stage, we provide systematic taxonomies and formally characterize the dependencies between adjacent stages, revealing how each stage both depends on and constrains the next. Beyond synthesizing existing work, we identify open challenges at stage boundaries and propose a cross-stage research agenda for closed-loop multi-agent systems capable of continuously diagnosing failures, reorganizing structures, and refining agent behaviors, extending current coordination frameworks toward more self-organizing forms of collective intelligence. By bridging these previously fragmented research threads, this survey aims to offer both a systematic reference and a conceptual roadmap toward autonomous, self-improving multi-agent intelligence.

A plethora real-world environments require agents to compete repeatedly for the same limited resource, calling for a temporal notion of fairness judged across entire interaction histories. This paper advances the theory of temporal fair division by introducing Rotational Periodicity (RP), a family of lightweight metrics, alongside the ALT family of sliding-window measures, within a unified framework for repeated multi-agent resource competition. We formalise the Multi-Agent Battle of the Exes (MBoE) as a repeated fair division instance and establish Perfect Alternation (PA) as its canonical temporally fair solution, drawing connections to proportionality, envy-freeness, and n-periodic round-robin allocation. RP decomposes temporal fairness into two complementary sub-measures: Rotational Score (RS) and Waiting Periods Evaluation (WPE), achieving O(nu+n) time complexity versus the O(nu*n) of ALT, where nu is the episode count and n the agent count. Empirical evaluation across n in {2,3,5,8,10} reveals three findings. First, both RP and ALT expose a coordination failure invisible to traditional metrics: Q-learning agents perform worse than random policies by 10-73% on RP and 7-35% on CALT, while Reward Fairness remains misleadingly high (above 0.92 for n>=3). Second, RP achieves 12-25x computational speedup over ALT, growing with n. Third, the two families are complementary: ALT provides richer discrimination for small populations; RP scales reliably where ALT becomes intractable. Together they form a diagnostic toolkit for temporal fair division.

Operational plan generation and verification are critical for modern complex and rapidly changing battlefield environments, yet traditional generation and verification methods still respectively face the challenges of generation infeasibility and verification insufficiency. To alleviate these limitations, we propose an Integrated Multi-Agent Framework for Generative Operational Planning and High-Fidelity Plan Verification (IFPV). IFPV consists of two tightly coupled modules: Multi-Perspective Hierarchical Agents (MPHA) for generative operational planning and an Adversarial Cognitive Simulation Engine (ACSE) for high-fidelity adversarial plan verification. MPHA decomposes commander intent into executable multi-platform tactical action sequences through the collaboration of Pathfinder, Analyst, and Planner agents. ACSE introduces an opponent equipped with a customized world model, which predicts the future evolution of mission-critical platforms and conducts dynamic counteractions against candidate plans. Simulation experiments in the Asymmetric Combat Tactic Simulator (ACTS) show that IFPV improves mission success by 19.4% and reduces operational cost by 41.7% compared with a single-step large language model (LLM) planning baseline. Compared with a traditional rule-based validator, ACSE increases the average suppression rate by 31.8%, indicating that the proposed verification environment is stricter and more discriminative in revealing the latent vulnerabilities of candidate plans. The code for IFPV can be found at https://github.com/zhigao3ks/IFPV.